Od. 16.305 reads as follows in West’s edition:Footnote 1
His critical note suggests that there are, in essence, two readings of the first hemiepes—καί κέ τεο δμώων and καί κ’ ἐτεοδμώων. These may be paraphrased, roughly, as ‘And we could also make trial of many a one of the slave men’ and ‘And we could also make trial of the men to see whether they are true slaves’. The former is championed by Herodian and found, with minor variations, in all principal manuscripts; the latter is advanced by Ptolemy of Ascalon and also attested by Hesychius.
In view of the overwhelming evidence of the manuscript tradition and of Herodian’s authority, West’s decision to adopt the vulgate reading as his text may seem secure. And yet the case is less clear-cut than it appears.
I
To begin with, West’s presentation of the former reading is not entirely accurate. Ω stands for the eight manuscripts on which the recension principally relies and, where pertinent, the assumed manuscripts of Eustathius and Tzetzes.Footnote 2 West’s parenthetical remark is, I take it, meant to explain the preceding ‘ita fere’: for in four of the eight manuscripts the pronoun is accented as an interrogative rather than an indefinite, while in G it is—or so it may appear—additionally misspelt with omega.
Closer inspection of Laur. 32.24 (G) shows that omicron and omega are scarcely, if ever, confused; iota adscript, by contrast, though usually written, is omitted several times per hundred lines.Footnote 3 Consequently τέω here likely represents not τέο but τέωι and is to be understood as a dativus modi (‘in which way?’). This is confirmed by the scribe’s treatment of the other two occurrences of τεωι in the Odyssey, both of which are likewise written without iota.Footnote 4 G, therefore, appears to preserve a distinct third reading.Footnote 5
Mon. gr. 519B (U), another member of West’s Ω, offers καί κέν τοι δμώων.Footnote 6 Again, the rhythm is unobjectionable and the sense straightforward. Other manuscripts outside Ω provide further variants.Footnote 7 Taken together, these readings suggest textual corruption: scribes seem to have struggled to make sense of what they read.
Much of this instability arose from a metrical anomaly: καί κέ τεο δμώων is an awkward sequence to scan. The present passage is the sole instance in the Iliad and Odyssey—indeed, in extant Greek epic—in which the indefinite pronoun τεο is unambiguously disyllabic: in all other occurrences, τεο either can or must be read monosyllabically. (The full data follow below.) No wonder, then, that some scribes—consciously or unconsciously—sought to regularize the line. The two variant readings in Ω are readily explained on that view: in G, τεο was replaced by the invariably disyllabic dative τεωι;Footnote 8 in U, the immediately preceding κε acquired a movable nu, while the infrequent τεο (for which medieval manuscripts elsewhere mostly transmit τευ) was replaced by the common emphatic τοι. The readings of G and U are thus best regarded as lectiones faciliores derived from καί κέ τεο δμώων.
Before setting out the data, I must briefly address a historico-linguistic question. In connection with these forms, West makes two slight but—because they recur more than two dozen times across the two epics—relatively conspicuous interventions in his editions:Footnote 9
First, West emends transmitted τευ to τεο, based on the observation that in early Ionic inscriptions the diphthong in words such as ἐκαλεῦντο was written εο. This orthographic principle has been generally (though not universally) accepted—at least in theory.Footnote 10 I shall not pursue the matter here, since the argument of the present paper does not depend on it: in all such occurrences the pronoun is indisputably monosyllabic, whether spelt τευ or τεο. The issue essentially concerns a slight change in vowel quality, with the corresponding change in spelling.Footnote 11
Secondly, whenever transmitted τευ stands in correption—when it occupies the first or second breve of a biceps and stands before a word beginning with a vowel—West restores not τεο but τε’, with elision of the final ο. An example in which the restored form occupies the second breve:
West’s argument for this second correction may be stated as follows. In word-final position, εο—unlike ευ—represents not a diphthong ([eo̯]) but two monophthongs in hiatus ([e.o]). Accordingly, when εο stands before a word beginning with a vowel, there can be no correption; instead, its final ο is elided before the following vowel. On West’s view, then, the disyllabic τεο in Od. 16.305, which is the subject of this section, is not isolated; it has more than a dozen parallels.
This second intervention has never received adequate scrutiny in the quarter-century since West’s Iliad was published; nor has it, so far as I can see, carried conviction.Footnote 13 The two principal problems are that West assumes without proof that τεο is disyllabic, and that τε’ for τεο is unattested—it first appears in A. Fick’s reconstruction of an Ur-Odyssey.Footnote 14
Anyone who accepts the first but not the second change should write τεο in either case. The choice is straightforward: it is one of the alternatives that West himself repeatedly mentions.Footnote 15 It is also the choice he made a few years earlier when editing Hesiod:Footnote 16
For the repertory of forms, this means that whenever transmitted τευ occupies a breve in the biceps before a vowel, I shall count it as a monosyllabic occurrence of τεο.
I now turn to the data. Three closely related forms must be considered: the indefinite τεο, the interrogative τέο and the relative and indirect interrogative ὅτεο, with its metrical variant ὅττεο:
(i) The indefinite τεο has twenty-one occurrences in the Iliad and Odyssey. One of them is the plainly disyllabic occurrence in καί κέ τεο δμώων (Od. 16.305). The remaining twenty are all either necessarily or possibly monosyllabic.Footnote 17 An example of a monosyllabic τεο in the princeps:
An instance in which a monosyllabic τεο occupies the first breve of the biceps before a vowel:
And an example of a possibly monosyllabic τεο in the biceps before a consonant:
(ii) The direct interrogative τέο occurs seven times in the two epics. Six occurrences are straightforward: in one instance the pronoun occupies the princeps and thus must be monosyllabic; in five it stands in the biceps before a consonant and is thus potentially monosyllabic.Footnote 18
One case remains in which transmitted τεῦ stands in correption before ἂν. On West’s principles, one would expect him to print (following Platt) τέ’ ἂν; instead, however, he deletes ἂν and prints (with van Leeuwen and Mendes da Costa) disyllabic τέο. Should that correction be justified, the disyllabic indefinite τεο at Od. 16.305—the focus of this paper—would no longer be isolated. The passage therefore requires closer examination. We begin with the paradosis:
Achilles awaits his mother’s return with the new armour forged by Hephaestus to replace that taken by Hector. Meanwhile Hector nearly succeeds in seizing Patroclus’ corpse. Hera intervenes by sending Iris to Achilles to urge him to prepare for battle. Achilles asks how he is to join the fight and then speaks the line quoted above.
The sentence is a syntactic blend—or, in the traditional terminology (preferred by WestFootnote 20 ), a contamination—of the following two sentences:
Achilles begins with (α) and then passes into (β): the former is a statement with an embedded relative clause, in which ὅτευ, the relative pronoun, refers back to ἄλλου τευ; the latter is a question, in which τεῦ is a direct interrogative pronoun.
The two sentences are inextricably intertwined: the change from ὅτευ to τεῦ is out of place in a relative construction; the indefinite τευ and the modal ἄν are out of place in an interrogative construction.
The blend must be intentional: had he wished, the poet could easily have expressed either sentence metrically. The assertoric sentence, for instance, might have run as follows:
The aim was to register Achilles’ initial bewilderment and his almost immediate resolve to fight.
So much for the transmitted sentence. As already noted, on West’s principles one would expect him to restore the older Ionic forms τεο and τέο and, in the latter case, to print τέ’ ἂν, since τέο here precedes ἂν and is taken by him to be disyllabic.Footnote 24 Instead, however, he deletes ἂν and emends the text to disyllabic τέο:
In his note on the passage, West first remarks that ‘the ἄν does not seem very much at home in a deliberative question’ and concludes that it is ‘probably better to expel the particle and write τέο κλυτά following Bechtel’.Footnote 25 Presumably West regards the expulsion of ἄν as preferable precisely because the particle is out of place in a question. There is no denying that ἄν appears odd in a question; yet this particle originates in the embedded relative clause of the source construction (α), and relative clauses in the subjunctive are often used with a modal particle.Footnote 26 More importantly, the deletion solves nothing: even without ἄν, West’s sentence remains syntactically ill formed, since the indefinite τεο too derives from the antecedent of the embedded relative clause in (α) and is therefore left unexplained.
The mention of Bechtel may suggest that the concern about ἄν in direct or indirect interrogatives goes back at least to that scholar.Footnote 27 In fact the relevant conjecture first appears in van Leeuwen and Mendes da Costa’s first Iliad: the editors deleted lines 192–5, but in the apparatus they recorded various possible corrections, among them disyllabic τέο for transmitted τεῦ ἂν. Those scholars’ concern was with ἄν tout court: it was an idée fixe of van Leeuwen’s in particular that Homer could only have used κε and consequently that all occurrences of ἄν must be either deleted or replaced by κε.Footnote 28 In this respect, then, Bechtel, van Leeuwen, etc. proceeded from an assumption now generally rejected. An appeal to their authority therefore carries little weight.
West was therefore not justified in deleting ἄν and introducing a dubious disyllabic τέο. Had he followed his own assumptions consistently, he should rather have emended transmitted τεῦ ἂν, with Platt, to τέ’ ἂν, just as elsewhere he emends prevocalic τευ in the biceps to τε’. In that case, however, the same two objections would still arise: the assumption that τέο—unlike transmitted τεῦ—is disyllabic is unproven; the apostrophized form τέ’ for τέο is unattested—it first occurs in Platt’s edition. Thus the seven occurrences of the direct interrogative τέο are all either necessarily or possibly monosyllabic.
(iii) The relative and indirect interrogatives ὅτεο and ὅττεο occur five times in all.Footnote 29 ὅτεο follows the trochaic caesura of the third foot, so that the sound corresponding to εο stands in the princeps and must therefore be monosyllabic; ὅττεο, by contrast, always occupies a whole foot before a consonant, so that the same sound may be monosyllabic.
The overall result may now be briefly restated. The occurrence of τεο in καί κέ τεο δμώων is metrically unique: it is unambiguously disyllabic, its first syllable belonging to the first foot, its second to the second. Some of the variant readings discussed at the outset may well have been prompted by this exceptional metrical shape.
There remains, finally, the question of sense. An initial, literal paraphrase of the line might be given as follows:
Odysseus presents Telemachus with his plan for how they are to proceed in their fight against the suitors (Od. 16.270). In particular, he intends to test the loyalty of the female slaves (16.304). He then adds the line quoted above, whose sense must clearly be ‘Let us also put the male slaves to the test’ (16.305). In his reply, Telemachus urges his father to reconsider the latter suggestion since the men are out in the fields—it would take too long to make trial of each individual (16.311–14). Regarding the women, however, he agrees that it is essential to determine who is and who is not guilty (16.316–17).
The difficulty with the line’s first hemistich lies in the sense of τεο. Translators and commentators have interpreted the pronoun in three different ways—as ‘some one’,Footnote 30 ‘many a one’,Footnote 31 or ‘every one’.Footnote 32
The literal meaning of ‘some one’ is evidently inapposite: as the context makes clear, Odysseus does not intend to make trial of any one of the male slaves; nor is there one particular man whom he wishes to test.
The meiotic use of ‘many a one’—the current favourite among interpreters—does not fare much better. Why should Odysseus wish to test only a subset of the male slaves?
The collective sense of ‘every one’ or ‘each one’ would satisfy the reader’s expectations from the context. Telemachus’ response shows that he understood Odysseus in this way, since he rephrases his proposal as ἑκάστου πειρητίζων (‘making trial of each’, Od. 16.313). In the case of the female slaves too, the implication undoubtedly is that Odysseus and Telemachus will test all women (16.304, 316–7). Whether τεο can bear this meaning is another matter: none of the examples adduced for collective τις is constructed with a partitive genitive.Footnote 33
The issue therefore remains unresolved, but the very fact that commentators disagree on the meaning of the word is itself significant, for it suggests that something may indeed be amiss.
To sum up, the vulgate reading is palaeographically unstable, metrically odd and semantically perplexing.
II
Were we to adopt the latter reading of West’s apparatus criticus, the line would read
The reading’s association with Ptolemy rests on the scholium ad loc.:Footnote 34
καί κέ τεο δμώων: τὸ τεο ἀόριστον· διὸ καὶ ἐπὶ τοῦ κέ ὁ τόνος. ὁ δὲ Ἀσκαλωνίτης ἐτεοδμώων, τῶν ἀγαθῶν θεραπόντων· τινὲς δὲ οὕτως ἐτεοδμώων, τῶν ὄντως ἡμετέρων δμώων. (Σ H Od. 16.305)
τὸ τεο Buttmann: τὸ τε H | ἐτεοδμώων2 Lentz: ὅτεο δμώων H | ὄντως Cobet: ὄντων H
A paraphrase: in καί κέ τεο δμώων—the vulgate reading just discussed—τεο is (not an interrogative but) an indefinite noun.Footnote 35 Being indefinite, τεο must be enclitic, meaning its tone inclines or leans on the immediately preceding κε, which should therefore be written with an acute accent. Yet the Ascalonite—Ptolemy of Ascalon—reads the line with ἐτεοδμώων, which he suggests should be understood as ‘of the good servants’. Other scholars, too, read ἐτεοδμώων in the present passage; but they interpret the word as ‘of the slaves that are really ours’.
The scholium has traditionally been assumed to have been drawn in toto from Herodian’s Odyssean Prosody.Footnote 36 This assumption is unlikely, however, as discussions of variant readings typically trace back to Didymus’ On the Aristarchean Recension.Footnote 37 More plausibly, then, the scholium is a composite: the initial section on the accentuation of τεο and κέ likely comes from Herodian, while the final section with the variant reading and its two interpretations stems from Didymus.Footnote 38
Ptolemy flourished in the late second century b.c.e.; Stephanus of Byzantium refers to him as a pupil (γνώριμος) of Aristarchus (α 476).Footnote 39 Didymus presumably recorded Ptolemy’s reading because it diverged from that of Aristarchus: hence what I earlier referred to as the vulgate reading—καί κέ τεο δμώων—is really the reading of Aristarchus.Footnote 40
The τινές who proposed the alternative interpretation remain otherwise unidentified. The plural form does not necessarily denote a multitude; indeed, given the proposal’s esoteric nature, τινές probably refers to a single individual.Footnote 41 That scholar likely advanced their interpretation in the late second or early first century b.c.e.—after Ptolemy but before Didymus.
A third interpretation of ἐτεοδμώων is attested in the lexicon preserved under Hesychius’ name:Footnote 42
According to the above gloss, ἐτεοδμώων means ‘of those serving as slaves very genuinely’.Footnote 43
There is reason to believe that Hesychius’ original lexicon was fully alphabetized.Footnote 44 The entry for ἐτεοδμώων thus likely represents a later interpolation, as the lemmata appear in the order Ἐτεοβουτάδαι, Ἐτεόκρητες, ἐτεόν, ἐτεοδμώων, ἑτεραλκέα δῆμον—in strict alphabetic sequence, ἐτεοδμώων should have been placed between Ἐτεοβουτάδαι and Ἐτεόκρητες. Nothing is known about the identity of the interpolator, nor about the date or source of the addition.Footnote 45
The ultimate source of the entry is also unknown. Its language, however, offers clues for establishing an approximate terminus post quem. The phrase γνησίως δουλεύειν reflects the concept of genuine slavery attested in the writings of the Church Fathers since at least Origen.Footnote 46 As Chrysostom observes, ‘a genuine slave is the slave of no one else’.Footnote 47 This patristic notion of exclusive servitude builds upon and sharpens a New Testament theme in which the faithful’s relation to God is compared to that of slaves to their master.Footnote 48 The gloss thus likely goes back to a Christian scholar of the third or fourth century c.e., though possibly later, who may have drawn on this theological idea in an allegorical reading of the Ithacan books—just as Odysseus tests the loyalty of his slaves on his return, so Christ judges the faithfulness of believers at his advent.Footnote 49
The three interpretations share their morphological and semantic presuppositions: ἐτεόδμως—more on the correct accentuation in a moment—is analysed as a compound of ἐτεός (‘true’) and δμώς (‘slave’), with its overall meaning presumed to be identical to the combined meaning of its constituent parts (‘true slave’).
Ptolemy’s interpretation is the simplest of the three: an ἐτεόδμως is taken to be an ἀγαθὸς θεράπων (‘good servant’). The compound’s two constituents are rendered verbum e verbo. Neither rendition is entirely satisfactory: one can easily imagine a person who satisfies one description but not the other. A Homeric example: the Iliad poet describes Patroclus as Achilles’ ἀγαθὸς θεράπων (Il. 16.165 = 17.388); he would never, however, characterize him as his ἐτεόδμως—though both a δμώς and a θεράπων engage in some kind of service, the latter is not considered another’s property.Footnote 50 Ptolemy’s proposal, then, amounts to little more than a rough adumbration.
The second interpretation retains the constituent δμώς in its paraphrase and renders ἐτεός as ὄντως ἡμέτερος (‘really ours’), thereby introducing a distinction between appearance and reality. A slave who belongs to Odysseus and Telemachus—who is theirs (in appearance)—may not really or truly be so; he may, for instance, have aided the suitors in ways that undermine his (apparent) masters’ interests. Consider Melanthius: de jure, he is ‘their’ slave; yet de facto—through his actions—he sides with the suitors, revealing that he is no longer ‘really theirs’.
The third interpretation accepts this division between appearance and reality but refines what is meant by the latter: an ἐτεόδμως is now defined as ὁ γνησίως δουλεύων (‘one serving as a slave genuinely’).Footnote 51 Imagine one of Melanthius’ unnamed brothers: his behaviour is immaculate, yet he loathes Odysseus and his family, longing for the suitors to annihilate their power. According to the second interpretation, he would pass Odysseus’ test and qualify as ἐτεόδμως—which intuitively seems incorrect. Hence the further refinement: for it is not sufficient to act in the right way; one must do so with the right mental state. The focus thus shifts from the external, objective world to the internal, emotional realm.
Collectively these three attempts—and probably others existedFootnote 52 —constitute evidence of a previously unappreciated fact: in antiquity, ἐτεοδμώων was regarded as an authentic manuscript reading, since otherwise there would have been no reason for scholars from the second century b.c.e. to the third or fourth century c.e. to propose competing interpretations of the word’s true meaning.
Apart from its occurrences in the Odyssey scholium and in the lexicon transmitted under Hesychius’ name, the word ἐτεόδμως is unattested. As the ancients have long recognized, it is indeed a compound: the first constituent is ἐτεός, in its combining form ἐτεο‑, as in Ἐτεόκρητες (Od. 19.176); the second is the monosyllable δμώς (e.g. Od. 24.257). This, in turn, settles the issue of accentuation—ἐτεόδμως or ἐτεοδμώς—since according to a widely attested rule that likely originates with Herodian, the word is paroxytone.Footnote 53 In Choeroboscus’ words:Footnote 54
πᾶν γὰρ ὄνομα μονοσύλλαβον ἐν τῇ συνθέσει ἀναβιβάζει τὸν τόνον. (Choer. In Thds. 1.186.38)
Every monosyllabic noun, when in composition, retracts its tone.
Further, ἐτεός derives from ἐτεϝός (/etewós/), as is evident, inter alia, from the Mycenaean form e-te-wo-ke-re-we-i-jo /etewoklewehios/ (PY An 654.8–9), a patronymic formed from what in a later script would be written Ἐτεϝοκλέϝης, the ancestor of Attic Ἐτεοκλῆς.Footnote 55 Originally, then, ἐτεός is trisyllabic (/e.te.os/); and in epic poetry, at any rate, its last two syllables are never contracted. From a metrical point of view, then, καί κ’ ἐτεοδμώων is impeccable.
As for the sense, the line as a whole may be paraphrased as follows:
And we could also make trial of the men to see whether they are true slaves.
On this interpretation, πειρηθῆναι takes a double-genitive construction: the first genitive (ἀνδρῶν) indicates the group being tested, as in the following πειράζειν (Od. 16.318–19); the second genitive (ἐτεοδμώων) specifies the standard against which they are tested.Footnote 56
In sum, the manuscript reading ἐτεοδμώων is ancient, metrically fine and semantically pertinent.
III
What is the aetiology of the two readings? Can either be explained in terms of the other—or perhaps both at once in terms of a tertium quid?
According to some, Ptolemy misread καί κέ τεο δμώων as καί κ’ ἐτεοδμώων; hence ἐτεόδμως, whose meaning he and later grammarians discussed, would prove to be a mere ghost word.Footnote 57 A psychological explanation for the corruption has also been proposed: Ptolemy must have been ‘influenced by the word Ἐτεόκρητες, which clung to his memory’.Footnote 58 According to others, however, ἐτεοδμώων must be a conscious conjecture of Ptolemy’s—an attempt to eschew τεο, ‘the unusual form of the genitive’ found in the transmitted text.Footnote 59
No one appears to have argued for a change in the opposite direction—from καί κ’ ἐτεοδμώων to καί κέ τεο δμώων. Which is slightly surprising, given that the former is clearly lectio difficilior and should, therefore, one might presume, be lectio potior.
Only two scholars, independently, have recognized that, in a certain sense, the original reading is not in question.Footnote 60 For in a papyrus (say) of the fourth century b.c.e., the hemiepes was—in either case—indisputably written as follows:Footnote 61
ΚΑΙΚΕΤΕΟΔΜΩΩΝ
In other words, the documentary evidence is not in doubt. What is uncertain is how to accent the transmitted text—how to parse and interpret the line. In this respect, Ptolemy’s variant stands on equal footing with the vulgate. True, not a single ancient or medieval Odyssey text has preserved the former. Yet that fact seems immaterial to establishing what the poet of the Odyssey originally intended. The two readings ought to be assessed chiefly on their inherent qualities; and if that is the decisive criterion, then as demonstrated in the previous sections, Ptolemy’s καί κ’ ἐτεοδμώων is preferable.
A parallel from the medieval transmission offers an instructive illustration of how such errors arose in antiquity. We have already encountered Ἐτεόκρητες: it, too, occurs only once in the extant remains of archaic epic poetry, at the same metrical location as ἐτεοδμώων—in the phrase ἐν δ’ Ἐτεόκρητες (Od. 19.176). Heid. Palat. gr. 45 (P), a member of West’s Ω, reads here ἐνδετέοκρῆτεσ.Footnote 62 The accentuation is significant not only as evidence of confusion arising from the Homeric hapax legomenon but also as a direct witness to the scribe’s attempt to reconstruct the verse beginning as ἐν δὲ τέο κρῆτες. Although the resulting phrase is syntactically incoherent, the scribal procedure mirrors that observed more than a millennium earlier in the case of καί κ’ ἐτεοδμώων and καί κέ τεο δμώων, offering a glimpse of an error, as it were, in statu nascendi.
Why, then, was ΚΕΤΕΟΔΜΩΩΝ initially analysed as κέ τεο δμώων and not as κ’ ἐτεοδμώων? The primary factor is undoubtedly that ἐτεοδμώων is a hapax. Confronted with this unfamiliar word, grammarians likely gravitated towards the more recognizable components of the alternative reading. The existence of superficially similar phrases such as κεν τοῦ (Od. 16.149) lent further support to this interpretation. Unlike Ἐτεοκρῆτες, where the misreading produced an ungrammatical phrase, the sequence in this case yielded a structure that, though erroneous, appeared prima facie coherent. Another factor is elision, which ‘presumably adds a degree of complexity to speech processing’.Footnote 63 While not the primary cause of misdivision, it may have compounded the difficulty of recognizing the correct segmentation.
Ptolemy is commonly thought to have discussed ἐτεοδμώων in his work on Homeric prosody, περὶ προσῳδίας Ὁμηρικῆς (Suda π 3038).Footnote 64 However, given that the issue at hand concerns not the accentuation or morphology of a single word but rather a variant reading and its interpretation, there is another, more likely possibility: Ptolemy’s book on Aristarchus’ editorial work in the Odyssey, περὶ τῆς ἐν Ὀδυσσείᾳ Ἀριστάρχου διορθώσεως (ibid.). If so, the passage quoted at the beginning of Section I—Σ H Od. 16.305—would constitute the first fragment that can be assigned to that work.
Based on the evidence discussed above, I suggest that Od. 16.305 be edited (in Westian style) as follows:
